Digital biomarkers for screening cognitive decline in vascular dementia

抄録

Toward the early intervention for dementia, there is a growing demand for methods to screen efficiently for cognitive decline. Neuropsychological tests based on interviews and paper-and-pencil tasks are widely used; however, challenges remain regarding the test-related burdens and the objectivity of assessments. In recent years, entirely new cognitive assessment methods using digital technology have been developed, and some have already reached the stage of clinical application. Specifically, this review outlines main digital modalities, including touchscreen-based testing, natural language processing-driven speech analysis, and eye-tracking-based assessments. These methods are expected to be applied as dementia screening tools that are marked by both convenience and objectivity. Particularly in the cases of vascular dementia or vascular cognitive impairment, frontal lobe dysfunction and attention deficits are observed in most cases, in which long-duration neuropsychological testing can impose a significant burden. Therefore, concise and objective assessments utilizing these specific digital modalities are highly relevant to clinical practice, as they alleviate the testing burden and are exceptionally well-suited for the early screening and longitudinal monitoring of vascular dementia. Finally, key challenges in artificial intelligence (AI)-based cognitive assessment, particularly the “black box” nature of AI and the critical need for diagnostic explainability, are discussed, along with future directions toward implementing cost-efficient, scalable screening methods.